On Wed, Aug 16, 2023 at 10:16:00AM +0200, Geert Uytterhoeven wrote:
> On Tue, Aug 15, 2023 at 6:50 PM Andy Shevchenko
> <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
> > On Tue, Aug 15, 2023 at 09:29:06AM +0200, Geert Uytterhoeven wrote:
> > > On Tue, Aug 15, 2023 at 9:19 AM Andy Shevchenko
> > > <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
> > > > On Sat, Aug 12, 2023 at 10:04:18AM +0100, Biju Das wrote:
...
> > > > > + {.compatible = "ti,dac5571", .data = &dac5571_spec[single_8bit] },
> > > > > + {.compatible = "ti,dac6571", .data = &dac5571_spec[single_10bit] },
> > > > > + {.compatible = "ti,dac7571", .data = &dac5571_spec[single_12bit] },
> > > > > + {.compatible = "ti,dac5574", .data = &dac5571_spec[quad_8bit] },
> > > > > + {.compatible = "ti,dac6574", .data = &dac5571_spec[quad_10bit] },
> > > > > + {.compatible = "ti,dac7574", .data = &dac5571_spec[quad_12bit] },
> > > > > + {.compatible = "ti,dac5573", .data = &dac5571_spec[quad_8bit] },
> > > > > + {.compatible = "ti,dac6573", .data = &dac5571_spec[quad_10bit] },
> > > > > + {.compatible = "ti,dac7573", .data = &dac5571_spec[quad_12bit] },
> > > > > + {.compatible = "ti,dac121c081", .data = &dac5571_spec[single_12bit] },
> > > >
> > > > I would reorder them a bit.
> > >
> > > Which is safe in this particular case...
> > > But not in general, as there might be fall-back compatible values.
> >
> > You mean the OF ID list must be specifically ordered?! What a nice minefield!
> > This has to be fixed somewhere else, surely.
>
> Seems like it is, cfr. the scoring system in drivers/of/base.c
> __of_device_is_compatible(). Sorry for the confusion.
>
> I still tend to order them in match tables though, from most-specific
> to least-specific.
>
> Note that soc_device_match() (which is used less, fortunately) does
> not have such a scoring system, so order does matter there.
Should be fixed there, because it's a big downside of OF.
--
With Best Regards,
Andy Shevchenko
